Skip to content
Crawler.sh Logo Crawler.sh Logo Crawler

Use Crawler MCP with OpenCode

OpenCode is an open-source terminal agent that speaks the Model Context Protocol. Once crawler-mcp is installed (guide), declare it in your opencode.json.

Pick a scope

Section titled “Pick a scope”
~/.config/opencode/opencode.json

Available to every directory you launch OpenCode from.

Add the server entry

Section titled “Add the server entry”

  1. Find the absolute path

    Section titled “Find the absolute path”
    which crawler-mcp

  2. Edit the config

    Section titled “Edit the config”
    {
      "$schema": "https://opencode.ai/config.json",
      "mcp": {
        "crawler-sh": {
          "type": "local",
          "command": ["/Users/you/.crawler/bin/crawler-mcp"],
          "enabled": true
        }
      }
    }

    Use the absolute path from step 1. OpenCode passes the binary the way it appears in the array, with no shell expansion.

  3. Restart OpenCode

    Section titled “Restart OpenCode”

    Close any running session and start a new one. The three tools (crawl_site, fetch_page, discover_links) show up under the crawler-sh server.

Raise the page cap

Section titled “Raise the page cap”

Set CRAWLER_TOKEN in the environment block to raise the per-crawl cap from 50 to 400 pages (or 10,000 with a Pro subscription). See Tier caps.

{
  "mcp": {
    "crawler-sh": {
      "type": "local",
      "command": ["/Users/you/.crawler/bin/crawler-mcp"],
      "enabled": true,
      "environment": { "CRAWLER_TOKEN": "your-token" }
    }
  }
}

Try it

Section titled “Try it”

In OpenCode, ask:

Use crawler-sh to fetch https://crawler.sh and summarise it.

OpenCode will call fetch_page and reason over the Markdown response.

See also

Section titled “See also”